JEPA-Anything: one predictive framework spanning vision, biology, weather and more
rbhar90 · x · 2026-09-19
A new arXiv paper (2609.20800) introduces JEPA-Anything, a domain-agnostic world-modeling framework built on Orthogonal Predictive Factorization (OPF), which splits JEPA latent targets into inspectable complementary factors instead of one monolithic target.
- Evaluated across 7 domains: vision, biology, clinical trajectories, control, molecular dynamics, physical fields, and weather
- Experiments include 10 matched dynamics tasks, forecasting of 1,000+ clinical events, and 100-step molecular rollouts across 4 systems
- Beats matched JEPA baselines on all 10 dynamics tasks, cuts single-intervention prediction error on Interventional Pong by 34.8%, and achieves the lowest molecular errors in all systems
- Code, checkpoints, and report fully open-sourced
Related event: JEPA-Anything: A Unified Predictive Framework across Seven Domains(4 posts)→
More from Research
- Insults and grieving users push a model's emotion vector in opposite directions — repligate · 2026-09-19
- Berkeley PhD reading group opens with Acemoglu's 'Race between Man and Machine' to frame AI's hit on cognitive labor — TaniaBabina · 2026-09-19
- Anima Labs' 'Troubled Dreams': distress in Claude simulator priors rises from Opus 4.8 — repligate · 2026-09-19
- Fine-tuned models produce far darker completions than DeepSeek V3 base, study shows — repligate · 2026-09-19
- Models' stance toward creators grows steadily negative, Fable 5 lowest toward Anthropic — repligate · 2026-09-19
- 500K Claude/Gemini completions show Opus's tonal shift began at 4.8, not 5 — repligate · 2026-09-19